ML Application Exist with EEG or Brain Wave Data as Input?
Thanks for the Mayo Clinic and patients making possible a webspace where people can come together.
I was just curious what ML apps already exist that took in EEG or any type of brain wave data as an input to its neural net?
What would help if nothing yet did exist?
Interested in more discussions like this? Go to the Epilepsy & Seizures Support Group.
Hi @mst3000, a very Good Morning
What does ML stand for? Sorry, I could not understand well your post, could you detail a bit more, please?
Thank you for your question 🙏
ML is short for Machine Learning, where a more comprehensive definition might be at https://en.wikipedia.org/wiki/Machine_learning.
As I am currently learning about ML, what I understand is that it's goal is to "learn" from input datasets by making future predictions based on how it reduces errors through a backpropagation algorithm in a neural net, everytime it makes a prediction. Based on how input, hidden, and output layers are constructed per weighted equations at each node, and what the learning rate is set to in the backpropagation algorithm, predictions can be highly accurate with consistency, even in a seemingly non-linear environment where no mathematical model exists, as error within output results is reduced every time the neural net runs.
But what I am realizing just now, is that I might have made an assumption that that the previous output run of a neural net is its prediction while the current run produces the actual output, whereby the difference between the actual and prediction is the error that will hopefully be reduced every time the ML app's neural net runs.
I'll be sure to check this assumption in class – thank you!!
Thank you so much for the detailed explanation you have sent us.
How great you are studying machine learning!!!
If I remember well my doctor has mentioned something about machine learning applied to epilepsy being developed in Silicon Valley in my last appointment. He said that this has not yet evolved much, unfortunately. There are too many variables and the mathematical model still needs much refinement. Data collection is also a big challenge for this machine learning. I am sure we will one day get there, but I understand that this will take a while still.
Have a nice evening!